I like what Mithran said and I think it follows a kind of thinking that’s good for analyzing where troops fit. It reminds me of some analysis in Magic: the Gathering.
The first thing that has to be dealt with is Gargantaur’s massive “setup cost”. 35 damage for 22 mana isn’t shabby but you can get more out of many other creatures. It scales up to 50, 65, 80, etc. Being realistic, I think expecting more than 2-3 deaths at activation time is unreasonable, so damage will usually be one of those 4 numbers, probably most often 50 or 65. It’s respectable. But you have to kill things to get there. Hmm.
Next, we want to look at “best case scenario analysis”. In this analysis, we look at a card’s best, average, and worst cases and think about how good they are vs. how often they might happen. Troops with frequent, powerful best cases or infrequent, still-good worst cases tend to be exciting. More often than not, the average case is what you get. Definitions work like that.
Gargantuar’s best case is a very strange game. The kind of shenanigans that makes him exceptionally powerful are if you can sac a creature that resurrects itself while also creating/exploding gems to get extra turns and refill the way you started this loop. I don’t think a loop with all of those qualities exists in-game, and if it does it’d probably take 4 troops to pull off so there’s no room for Gargantuar. Alternately: if you could find a way to consistently damage then heal Gargantuar you can boost him that way. Again, if you aren’t generating turns it’s not going to work, so I struggle to think of a troop combination that can do it. So the most likely best-case is when Gargantuar is your last troop with full mana vs. the enemy’s last troop with no options. You can activate it for a whopping 125 damage to win the game. But few troops are bad in that position.
The worst case seems like what happens when trying too hard to hit the best case fails. Let’s disqualify some obvious worst cases where almost every card fails and instead focus on some tight scrapes where cards can dig back out. For example, let’s say the board gets whittled down to 1v2, Gargantuar staring down 2 opponents. I think it does “average” or “below-average” here. It can hit for 110 damage if activated, but raising 22 mana by itself isn’t going to happen fast. Even so, there’s still the 2nd enemy to deal with. Sure, that third trait boost it as it takes damage, but how many hits is Gargantuar really going to take at that point? 2? 3? Once you fall behind, Gargantuar gets worse the further behind you get. Lame. I’ve come back from this via Crimson Bat or TDS many times, but I’m not sure how Gargantuar wins in that situation.
I think the average case is Gargantaur does 2 activations, one for 50 and one for 65 damage. It’s a decent sniper for a looping/exploding team, but I wonder if you wouldn’t want to just keep your loops/explosions going.
So same TL;DR:
Gargantuar’s best case happens when you start out by trying to lose the game. That best case isn’t a very good way to win a game where you’re behind. “I’m the best when you’re winning anyway” isn’t exciting. The worst case is the same worst case any card generally has, and the average case feels out-of-place for the kinds of teams it might fit on. If we get some really weird new other mythics that enable infinite self-harm loops, it’ll be at home on those teams. Without that, it’s good for a late-game snipe but is a very bad last troop standing. If it had any kind of resurrection chance via traits, this would get REALLY interesting.
It’s a troop that’s only fantastic when you’re losing a little, but not by a lot. It turns that into a winning game. Find a troop that can turn “losing by a lot” into a winning game instead, or just run a team that doesn’t have to start losing to win. Needs support troops that don’t exist and are too broken to be developed.